jwitchel

I knew a guy early in my career who has a secretary (for starters!) who printed out his email and put them on his desk every morning with the rest of the mail. He would then dictate his response to her and she would type it in. It was pure ego. Nothing else.

There is near zero actual risk. These systems have been around for a long time and are widely deployed in Europe.

Circuit breakers have been around for a century and work great. You plug something in that pushes or pulls too much current and pop - disconnect.

So while they are absolutely safe, there's huge FUD coming from the utilities who lose revenue AND PROFIT they were getting from you with each system.

Pacific Gas and Electric, Southern California Edison, Florida Power & Light (FPL), Commonwealth Edison (ComEd), Consolidated Edison (ConEd) are all for profit monopolies.

Asking them if it's safe is like asking Microsoft if you should switch away from Word. It does not matter the reasoning, the answer was always going to be the same.

"I'm gonna have to science the shit out of this."

Fantastic work!

I was looking for a dead on feature for feature exact replica including look and feel. You are 100% correct there are some excellent tools and libraries that duplicate most (all?) of the features. But I was playing around with the idea of what it would take for a perfect replica (for science of course). Turns out it's really hard even if you jump start it with those libraries and an existing reference implementation.

That was the point of my comment... Robust commercial software is hard.

I recently asked Claude to duplicate Google Docs. It practically threw up. I tried to have it write a plan, decompose it, deobfuscate it.

It gave me all sorts of reasons why this was a terrible idea. I've never seen it resist a task so directly and relentlessly.

It knew.

One point worth considering is that tools like Jira and Salesforce have dozens of screens and modals. But you only ever look at one at a time. So the enormity of the ask "duplicate Jira" is hard to see in its totality.

With Google docs, the entirety of the tool is almost one screen. It resists decomposition. So the true gravity of the request is more in your face.

You say you want to duplicate Asana or Service Now or Jira or Zendesk? Great, here's the keys to the car, a tank of gas, and a quarter to call me on a payphone when you get there. Oh wait payphones don't exist anymore...but it doesn't matter because you're never getting there.

These software platforms are built by thousands of engineers over more than a decade of dedicated work. They are they way they are for reasons. To think someone can duplicate them with some clever prompting is to completely fail to understand the scale of the problem at hand.

Kagi is awesome. And frankly, it's the end game. Just pay for your search full stop. The original 1990's assertion that "information wants to be free" and all that related nonsense never really passed the laugh test.

Being able to search the web is critical to day-to-day life. Why wouldn't you pay for it? I pay for all sorts of dumb stuff I barely use. I use Kagi every day.

Because I am paying for their service, I know that they are trying to provide me the best service they can - we are aligned. I want to pay for good service, they want to get me good service.

I pay for Kagi not because of political value system. I pay for it because it's just better than Google. I encourage you to try it. It's night and day.

Oh this is one I can answer. I was the co-founder of Prosper an early p2p lending company. Your problem is a relatively easy one.

Go find one or two companies with a lot of packages. Focus on delivering packages whose delivery time is not critical (say 7 days) and whose registered value is low (say $500 or less). Focus on deliveries that are just outside of ubers practical range (say LA to SFO).

Find a bunch of uber drivers who are already reliable. Over pay them to make runs. If they screw it up have them drop the package off at a ups store and over night it as a catch all.

Do 1,000 packages to get your logistics worked out. Don't market or launch or promote until you have it nailed.

Then find 10 more companies with lots of packages. Your goal is to develop profitable routes (e.g. 5 packages per car) and customers who are delivering 1000s of packages a month.

Your pmf is about reliably servicing one geographic region at a time like yelp and LinkedIn did.

Do not try and roll out a national service day one.

Do not focus on people shipping packages. Not enough repeat business.

Do not take high stakes runs day one.

One big advantage is your ability to do small heavy objects. Ups and FedEx are very weight sensitive.

This is a great incredibly well written piece. Nice work showing under the hood build up of how a db works. It makes you think.

Look a rural electric coops like www.lpea.coop if you want a battle tested approach to an org structure that resists the inescapable profit dynamics of a corporation.

AI slop for all its banality may give us enough noise in the signal to accomplish exactly what the author is asking for.

As the dead web continues to emerge, content looks less like apples on a tree and more like sand on the beach.

And the act of looking for a misshapen grain of sand becomes absurd.

Been using Kagi for about a year (paid). Best money I ever spent. I did a google search recently... Yuck.

I want a calm internet. I ask it answers. No motive. No agenda. Just a best effort honest answer.

I've been using Kagi search for a while now and frankly it's fantastic. Google looks like AOL to me now.

These guys are doing great work and this news product is exactly what I want... Once a day hit. What is happening in the world? As far as pmf goes they hit the mark for an old fart like me.

Usually the observation that “a widely used thing is objectively bad” is strong market signal for entrepreneurial opportunity in a big tam.

I for one would welcome a set of deeply integrated ui improvements in a Mac that included a better file manager, better window management, better desktop search, a contact manager just that worked, a messaging client that just worked, audio and camera controls that just worked, a calculator that didn’t suck, etc.

I’d pay at least $100 a year for that tool set.

And the inevitable despair of being smart in Milwaukee.